Article Text

Download PDFPDF

Spastin mutations are frequent in sporadic spastic paraparesis and their spectrum is different from that observed in familial cases
  1. C Depienne1,
  2. C Tallaksen1,
  3. J Y Lephay1,
  4. B Bricka2,
  5. S Poea-Guyon1,
  6. B Fontaine3,
  7. P Labauge4,
  8. A Brice1,
  9. A Durr1
  1. 1INSERM U679, Groupe Hospitalier Pitié-Salpêtrière, Paris, France
  2. 2Unité de Neurogénétique, Département de Génétique, Cytogénétique et Embryologie, AP-HP, Groupe Hospitalier Pitié-Salpêtrière
  3. 3Fédération de Neurologie and INSERM U546, Groupe Hospitalier Pitié-Salpêtrière
  4. 4Service de Neurologie, CHU Montpellier-Nîmes, Hôpital Caremeau, Nîmes, France
  1. Correspondence to:
 Professeur A Brice
 INSERM U679, Groupe Hospitalier Pitié-Salpêtrière, 47 boulevard de l’hôpital, 75013 Paris, France; brice{at}


Background:SPG4 encodes spastin, a member of the AAA protein family, and is the major gene responsible for autosomal dominant spastic paraplegia. It accounts for 10–40% of families with pure (or eventually complicated) hereditary spastic paraparesis (HSP).

Objective: To assess the frequency of SPG4 mutation in patients with spastic paraplegia but without family histories.

Methods: 146 mostly European probands with progressive spastic paraplegia were studied (103 with pure spastic paraplegia and 43 with additional features). Major neurological causes of paraplegia were excluded. None had a family history of paraplegia. DNA was screened by DHPLC for mutations in the 17 coding exons of the SPG4 gene. Sequence variants were characterised by direct sequencing. A panel of 600 control chromosomes was used to rule out polymorphisms.

Results: The overall rate of mutations was 12%; 19 different mutations were identified in 18 patients, 13 of which were novel. In one family, where both parents were examined and found to be normal, the mutation was transmitted by the asymptomatic mother, indicating reduced penetrance. The parents of other patients were not available for analysis but were reported to be normal. There was no evidence for de novo mutations. The mutations found in these apparently isolated patients were mostly of the missense type and tended to be associated with a less severe phenotype than previously described in patients with inherited mutations.

Conclusions: : The unexpected presence of SPG4 gene mutations in patients with sporadic spastic paraplegia suggests that gene testing should be done in individuals with pure or complicated spastic paraplegia without family histories.

  • AAA, ATPases associated with diverse cellular activities
  • DHPLC, denaturing high performance liquid chromatography
  • HSP, hereditary spastic paraparesis
  • SP, spastic paraparesis
  • SPG4
  • mutation
  • spastic paraplegia

Statistics from

Hereditary spastic paraplegias (HSP) are a genetically heterogeneous group of neurodegenerative disorders clinically characterised by progressive stiffness and weakness of the lower limbs. These symptoms are the result of axonal neurodegeneration of the cortico-spinal tract. Pure and complicated forms of HSP have been described, depending on whether spasticity occurs in the absence or the presence of other clinical features such as cerebellar ataxia, neuropathy, retinal degeneration, cognitive impairment, dementia, or epilepsy.

The clinical heterogeneity of HSP is partly explained by a large genetic heterogeneity. To date, at least 26 loci have been identified, associated with autosomal dominant, autosomal recessive, and X linked modes of inheritance. The most common form of autosomal dominant HSP (AD-HSP) is caused by mutations in the SPG4 gene, encoding spastin, a member of the AAA protein family (AAA, ATPases associated with diverse cellular activities). More than 150 mutations have been identified all along the SPG4 gene, including missense, nonsense, and splice site mutations, as well as frameshifts and larger deletions. SPG4 has been shown to account for 15–40% of all autosomal dominant HSP families, depending on the population.1–4 Emerging evidence suggests a role for spastin in microtubule dynamics,5–9 but the mechanisms by which spastin abnormalities lead to axonal degeneration remain largely unknown.

Based on the observation that many mutations reduce the abundance of the normal full length transcript or functionally normal spastin protein, the pathogenic mechanism is likely to be haploinsufficiency—that is, loss of function rather than a dominant negative effect. HSP caused by SPG4 mutations is generally described as a pure form of the disease—that is, as spastic paraparesis often associated with decreased vibration sense in the lower limbs and urinary problems. There is evidence that, at least in some cases, SPG4 mutation carriers express a more complicated phenotype associated with cognitive impairment,10–13 or in very few cases with cerebellar signs or epilepsy.14–18 Age at onset is highly variable, ranging from early infancy up to the eighth decade.

In addition to age dependent penetrance, estimated at 85% by age 45, there are asymptomatic carriers who are still completely normal or have abnormal signs only when examined, even at 76 years of age.1 To date, genetic testing for SPG4 remains limited to patients with familial histories, consistent with autosomal dominant or at least dominant transmission. In contrast, the frequency of SPG4 mutation carriers has not been systematically evaluated among isolated cases without family histories, and the question of de novo mutations has not been addressed. In the present study, we screened the SPG4 gene in 146 isolated patients by denaturing high performance liquid chromatography (DHPLC), to determine the frequency of SPG4 mutations among patients without family histories.



We included 146 cases with pyramidal signs present predominantly in the lower limbs who were referred by neurologists after exclusion of neurological causes such as multiple sclerosis, other leucodystrophies, intramedullar tumours, primary lateral sclerosis, or adrenoleucodystrophy when appropriate. This allowed us to include probands presenting either pure (n = 103) or complicated (n = 43) spastic paraplegia. The complicated forms included cerebellar signs or atrophy on cerebral imaging (n = 15), signs of peripheral neuropathy on electromyography and conduction velocity studies (n = 14), cognitive impairment (n = 3) and mental retardation (n = 5) on neuropsychological testing, ophthalmoplegia (n = 3), or tremor n = (3).

We determined the absence of family history of the disease for each patient by systematic interview of the probands on their first and second degree relatives. In 18 cases, there was doubt as to a possible gait disturbance in first degree (n = 12), second degree, or even more distant relatives (n = 6), but information from the proband was insufficient to conclude that there was a familial spastic paraparesis. As the relatives were deceased or not available for examination, these cases were classified clinically as sporadic. Both parents were examined and were found normal in 11 families. In six families, only one parent was available for clinical examination, and in 28 cases, at least one parent had died before the age of 50. The patients were mostly of European origin (France (122), Italy (4), Portugal (3), Spain (2), Greece (1), Poland (1)), but 11 patients were from other geographical regions (North Africa (4), elsewhere in Africa (2), West Indies (1), Madagascar (1), Turkey (2), Asia (1)). Two had been adopted. In addition, 300 European and 100 North African controls (mostly healthy spouses of patients with other neurological diseases) were included to test new variants of the spastin gene.

Informed written consent was obtained from each individual before blood sampling. This study was approved by the ethics committee Paris-Necker (CCPPRB No 03-12-07, 2/10/2004).

DHPLC screening

To evaluate the frequency of SPG4 patients in isolated cases with spastic paraplegia, we analysed the 17 coding exons of the SPG4 gene by denaturing high performance liquid chromatography (DHPLC). Experimental conditions for DHPLC mutational analysis of the coding region of the spastin gene were first set up with DNA from 30 patients with well characterised mutations in each of the SPG4 exons. All mutations were detected by DHPLC at the temperatures selected for the analysis.

The whole coding region of the SPG4 gene was amplified by polymerase chain reaction (PCR), using 18 primer pairs. Formation of hetero-duplexes was enabled before DHPLC analysis by denaturation (five minutes at 95°C) followed by gradually cooling to 25°C. DHPLC analysis was carried out at a flow rate of 1.5 ml/min for a time period of 2.5 minutes on a WAVE DNA fragment analysis system HSM 3500HT (Transgenomic, Omaha, Nebraska, USA). The temperature of the column was set to the exon specific melting temperatures for successful resolution of hetero-duplexes. Wild-type samples were always used as negative controls, to ensure that a normal homo-duplex profile was reproducibly obtained with regard to retention time and peak profile. Chromatograms from each patient were overlaid with one from a normal individual. Samples with extra peaks or with a difference in peak appearance were scored as positive. Primers and DHPLC conditions are listed in table 1.

Table 1

 Primers, polymerase chain reaction and denaturing high performance liquid chromatography conditions

Sequence analysis

Samples showing abnormal elution profiles were re-amplified from genomic DNA. Both forward and reverse sequence reactions were done using the Big Dye Terminator cycle sequencing ready reaction kit (PE Applied Biosystems, Foster City, California, USA). The sequence products were analysed on an ABI 3730 automated sequencer (PE Applied Biosystems).

Statistical tests

Frequencies were compared with the χ2 test or Fisher’s exact test when appropriate. Quantitative variables were compared by analysis of variance (ANOVA). Statistical analysis was done using SPSS software.


We identified 19 SPG4 mutations at the heterozygous state in 18 of the 146 sporadic cases with spastic paraplegia, for a frequency as high as 12% (table 2).

Table 2

SPG4 mutations in patients with sporadic spastic paraplegia

The frequency of SPG4 mutations was higher (16/103, 15.5%) in patients with a pure phenotype than in those with a complicated form of the disease (2/43, 4.6%) (p = 0.1). Both the latter had cognitive impairment (table 3B). Ages at onset ranged from childhood up to 70 years, with a mean (SD) age of 28.6 (17.9) years (excluding two patients with onset in childhood at an undetermined age). The overall severity was moderate after a mean duration of the disease of 17.4 (11.9) years (table 3A). One patient used a wheelchair after a disease duration of about 60 years. Interestingly, she had a complicated form of HSP (table 3B).

Table 3

 Clinical and familial characteristics of sporadic spastin mutation carriers with (A) pure and (B) complicated hereditary spastic paraparesis

We identified 14 missense mutations (74%), three splice site mutations (16%), one nonsense mutation (5%), and one frameshift (5%) (table 2). Four of the missense mutations (p.Gly370Arg, p.Arg424Gly, p.Arg460Cys, and p.Arg499Cys) have been described previously in AD-HSP families in which they were shown to segregate with the disease.1,6 All other mutations were novel. In order to exclude the possibility that these novel heterozygous variations were in fact rare polymorphisms, we screened a healthy white population of 600 chromosomes. None of the variations was detected in the controls or previously described in patients, suggesting that they are likely to be involved in the HSP phenotype. The mutation c.415+1 g→a, identified in a patient from North Africa, was also absent from 200 chromosomes of North African controls. The c.1348_1352del5 and p.Arg581X mutations are both presumed to lead to a premature stop codon and a truncated protein, compatible with a deleterious effect and loss of spastin function. The p.Pro238Thr, p.Gly385Trp, p.Ile406Val, p.Asp444Glu, p.Leu461Pro, p.Arg499His, p.Arg503Trp, and p.Asn579His mutations modify highly conserved amino acids in mouse, rat, and pig spastin orthologues. While p.Pro238Thr is in a domain conserved only in mammals, all other mutations modified amino acids that are also conserved in fish and arthropods, suggesting that they are essential for spastin function.

Interestingly, one patient had two different missense mutations (p.Ile406Val and p.Asn579His) that were not found in any other patient. Unfortunately, only DNA of the patient was available, so we were unable to determine whether these mutations were inherited in cis (on the same allele) or trans (each on a different allele). If the mutations were inherited in trans, the association of these mutations could be responsible for the disease and therefore for its isolated nature in this patient. Another possibility is that only one of these mutations is causative whereas the other is only a very rare polymorphism.

In addition to the 19 SPG4 mutations, we found one synonymous base change (c.390C→A/p.Ala130Ala, patient 39) and four different intronic variations close to an exon-intron boundary in six patients (c.1494-3dupt, c.1728+32c→g, c.1729-20t→a, and c.1496+18g→t in two patients; table 4). They were not detected on 600 control chromosomes, but until it is proven that they have an effect on the SPG4 expression or splicing we consider them to be non-pathogenic.

Table 4

SPG4 variations with unknown effect and polymorphisms in patients with sporadic spastic paraplegia

In addition to the 25 variants not found in controls, we identified heterozygous variants in seven patients which were also found in the controls (tables 2 and 4). In four patients, a previously described synonymous polymorphism in exon 6 (c.879G→A/p.Pro293Pro) was detected. In the remaining three, a non-synonymous polymorphism, c.131C→T/p.Ser44Leu, was found in exon 1. The frequencies of c.879G→A/p.Pro293Pro and c.131C→T/p.Ser44Leu were similar in the HSP patients (A879/Pro293: 3/146 = 2%; T131/Leu44: 3/146 = 2%) and in the control population (A879/Pro293: 10/300 = 3.3%; T131/Leu44: 8/300 = 2.7%). The p.Ser44Leu allele was first described as a causal mutation at the homozygous state,6 but was recently shown to be a polymorphism present in the general population which might act as a modifier of the HSP phenotype, especially the age of onset, when associated with a mutation.19 Two of the carriers of these polymorphisms also had causative mutations: patient 048 had the c.879G→A/p.Pro293Pro polymorphism associated with the p.Asp444Glu mutation, and patient 453 had the non-synonymous p.Ser44Leu polymorphism combined with a p.Arg499His mutation. In the latter, the age at onset was very early (childhood) and the disease severe after 40 years of evolution.

We then examined whether our patients had apparently isolated disease because of insufficient genetic or clinical data (for example, the early death of parents, adoption, or small families). Eighteen of the probands reported a very questionable gait disturbance in another family member who was deceased or unavailable for examination. As information given by the proband was insufficient to formally conclude that the disease was familial, these patients were classified as sporadic (table 5). Five mutations were identified in these 18 patients (28%), three of which were second degree relatives or even more distantly related. This suggests that even weak evidence for a secondary case with HSP should be taken into account for molecular testing. Five mutation carriers (15%) were found in the 34 patients in whom one parent had died before the age of 50 (n = 28) or was unavailable for clinical examination (n = 6). More surprisingly, we identified one mutation (9%) among the 11 patients whose parents were both normal on clinical examination (table 5). Analysis of parental status in this family (008) showed that the p.Arg581X mutation was transmitted by the mother who was asymptomatic at age 67, excluding the possibility of a de novo event.

Table 5

 Frequency of SPG4 mutations according to the familial data

We were intrigued by the large proportion of missense mutations in our series compared with previously studied European populations with AD-HSP, in which missense mutations represented approximately 30% of all SPG4 mutations and were generally clustered in the AAA cassette.1 In our study, 74% (14/19) of the mutations were missense, two of which (p.Glu43Asp and p.Pro238Thr) were located outside the AAA cassette. In order to compare the frequencies of missense mutations versus truncating mutations, we pooled the published data and our own unpublished data. Missense mutations were more frequent among sporadic cases (72%, 13/18) than in familial HSP (31%, 49/160), and this difference was statistically significant (p = 0.001). In contrast, truncating mutations—found throughout the coding region of the gene—represent approximately 50% of the mutations in AD-HSP,1 but were found in only two (11%) of our cases (c.1348_1352del5 and p.Arg581X). Furthermore, with the exception of the c.1348_1352del5 frameshift, all the non-missense mutations identified in our isolated probands were surprisingly uncommon. Two of the splice site mutations we identified are novel, and both affect the donor site next to exon 1. The p.Arg581X nonsense mutation was located at the C-terminus of the protein. As a milder phenotype or a lower penetrance could account for the isolated nature of our patients, we tested the hypothesis that mutations identified in this study might be associated with less severe disease. We compared 64 probands from SPG4 AD-HSP families and 18 sporadic SPG4 mutations carriers. Sporadic SPG4 mutations carriers were less severely affected than SPG4 familial HSP patients (40.6% of AD-HSP patients needed help with walking, compared with only 16.7% of sporadic patients; p = 0.060), although the mean (SD) age at onset in familial (28.6 (16) years) and sporadic cases (28.7 (17) years) was similar, as were the mean disease duration and the mean age at examination. Taken together, these results indicate that in isolated patients, SPG4 mutations are associated with a milder spastic paraparesis phenotype (and potentially reduced penetrance), which is responsible for apparently sporadic HSP.


Molecular analysis of the SPG4 gene in this series of 146 patients with spastic paraplegia but without family histories, and after exclusion of other major neurological causes, identified at least 18 mutations carriers. The proportion of SPG4 mutation carriers who appear to be isolated cases is therefore 12%. Mutation screening was carried out using DHPLC, which has already been used by other groups successfully to screen mutations in the spastin gene.20,21 Although this method is efficient and reliable, its sensitivity is not complete, and screening of the coding sequence only may miss mutations in intronic or regulatory sequences or large scale rearrangements.

Seventeen of the 25 different DNA variants identified were novel and none was found on 600 control chromosomes. The causative role of 18 of the 25 mutations has already been demonstrated (n = 5) or is highly probable because they are predicted to produce truncated spastin (n = 2) or spastin with a missense affecting a conserved amino acid (n = 9) or to alter splicing (n = 2). The remaining DNA changes were a synonymous base change and four intronic variants which could affect spastin mRNA splicing, but their consequences remain to be determined in reverse transcriptase PCR experiments.

Stratification according to available information on the family reveals that the frequency of spastin mutations carriers is greatest (5/18, 28%) when there is even a weak suspicion of a secondary case in the family. If the doubtful cases are excluded, the proportion is still 10% (13/128), suggesting that molecular diagnosis of SPG4 is indicated for patients with sporadic spastic paraplegia, after exclusion of other frequent causes of spasticity. The observation of a higher frequency of spastin mutations in isolated cases with pure spastic paraplegia (16/103) than in those with complex disease (2/43) is a potentially useful indication for molecular diagnosis.

The apparent isolation of the patients we studied may have several explanations—for example, lack of clinical data for the family and censor effects because of the early death of a parent. In everyday clinical practice, the parents and distant relatives are rarely available for examination and sampling, especially when the clinical interview does not suggest a genetic basis for the disease. Furthermore, although de novo SPG4 mutations are possible, none has ever been identified. Mutations are therefore probably transmitted and the apparent sporadic nature of the patients results from reduced penetrance, as in family 8. Normal clinical examinations of both parents is therefore insufficient to exclude the possibility of an SPG4 mutation.

Our results indicate that missense mutations are more common in isolated than in familial cases and are associated with milder disease. Genotype–phenotype correlations have not been very informative so far1; however, penetrance was not analysed according to the nature and the location of the mutations in the SPG4 gene and some of the mutations in sporadic cases might have a lower penetrance than mutations identified in families. Furthermore, in most studies, the theoretical mutation type was used for correlations studies, which is not very informative as presumed “missense” mutations may differ in their effects at the mRNA and protein level. The selection of apparently isolated patients in this study could thus have facilitated the detection of missense mutations associated with reduced penetrance and disease severity.

Our study supports the existence of negative severity modifiers, such as the p.Ser44Leu allele, which is also present in the control population. Although p.Ser44Leu is not a susceptibility factor for spastic paraparesis because its frequency is similar in HSP patients and controls, it could contribute, in association with an SPG4 mutation, to an earlier age at onset.19 Four different families with a mutation associated with the p.Ser44Leu allele have been described in the literature on SPG4, and p.Ser44Leu in the homozygous state was reported to cause spastic paraparesis.6,19,22 We found one patient (453) with this polymorphism and the p.Arg499His mutation who had an early age at onset as well as severe spasticity. We also identified one patient (113) with two different missense mutations (p.Ile406Val and p.Asn579His), who had severe spasticity. These two patients, of 146, suggest that the combination of the two variations could cause a form of the disease that is more severe than in other family members. This also raises the possibility that other intronic 5′UTR or regulatory sequence variations that were not detected by DHPLC might affect the expression of SPG4 and modify the clinical phenotype. Other genetic, epigenetic, or environmental factors could also affect disease expression in sporadic SPG4 mutation carriers. Dominant genes, such as SPG3A, might be involved as well, although no sporadic cases with this mutation have been reported.

These results have implications for genetic counselling, as the risk for offspring of sporadic patients with SPG4 mutations is difficult to evaluate. In a few cases it is difficult to determine completely the pathological nature of the variants identified, and the interpretation of the molecular data has to be handled with care in clinical practice. In most cases, the mutation is inherited in a dominant manner, but is associated with reduced penetrance. However, the factors determining the clinical expression and severity of the disease are not yet known. Nevertheless, testing for SPG4 mutations in patients with pure spastic paraplegia, even in the absence of a positive family history, would not only alert offspring of sporadic patients to the potential risk of HSP but would also help identify these factors.


We are very grateful to Dr Merle Ruberg for critical reading of the manuscript, to Isabelle Lagroua for examining the patients’ files, and to Dr Filippo Santorelli for DHPLC conditions of exons 1, 13, and 15. Many thanks to the families for participating, to the SPATAX network, and to the clinicians who referred their patients: Drs Jean-Pierre Azulay, Gerard Ponsot, Jean-Michel Vallat, Gerald Rancurel, Jean Pierre Decroix, Charles Pierrot-Deseilligny, Paula Coutinho, Diana Rodriguez, Marc Verny, Christine Tranchant, and Myriem Abada-Bendib. This work was supported financially by the VERUM foundation and the Programme Hospitalier de Recherche Clinique AP-HP (No AOM03059, to AD).



  • Published Online First 31 July 2005

  • Conflicts of interest: none declared

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.